Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Unrecoverable
Data
The Need for Drive-Independent Data Recovery
Commissioned by
ActionFront
Data Recovery Labs, Inc.
Written by
Charles H. Sobey
April 14, 2004
Table of Contents
1. Executive Summary............................................................................................................... 3
2. Introduction to Hard Disk Drive (HDD) Technology ........................................................... 4
2.1 Areal Density and Price Trends....................................................................................... 4
2.2 What Happens to Data in a Hard Disk Drive?................................................................. 5
2.2.1 Organizing the Data.................................................................................................. 5
2.2.2 Locating the Data ..................................................................................................... 7
2.2.3 Detecting the Data .................................................................................................... 8
2.2.4 Decoding the Data .................................................................................................... 9
2.2.5 Drive Burn-in and Optimization: Hyper-Tuning................................................. 11
3. Data Recovery Market......................................................................................................... 12
3.1 Perception vs. Reality .................................................................................................... 13
3.2 A Call for Transparency ................................................................................................ 13
4. Data Recovery Technology ................................................................................................. 14
4.1 Traditional Hardware Replacement Methods ................................................................ 15
4.1.1 Replace the PCB..................................................................................................... 15
4.1.2 Replace the Firmware............................................................................................. 16
4.1.3 Replace the Head Stack .......................................................................................... 18
4.1.4 Move the Disks to Another Drive........................................................................... 18
4.2 Magic Machines and Proprietary Processes ........................................................... 18
4.2.1 Spin-Stand Testers.................................................................................................. 19
4.2.2 Magnetic Force Microscopes (MFM)..................................................................... 20
4.2.3 The Spin-Stand MFM? ........................................................................................... 20
4.2.4 Exotic Recovery ..................................................................................................... 21
5. The Frontiers of Whats Possible: What Makes Data Unrecoverable?................................ 21
5.1 When Firmware Replacement Fails............................................................................... 21
5.2 When Head Stack Replacement Fails............................................................................ 22
5.3 When Disk Remounting Fails........................................................................................ 22
5.4 When the DATA Fails................................................................................................... 23
6. Future Success Depends upon Developing Drive-Independent Data Recovery Capabilities
................................................................................................................................................. 24
7. The FIRST Public Demonstration of Drive-Independent Data Recovery: ActionFronts
SignalTraceTM Technology...................................................................................................... 25
8. Conclusions ......................................................................................................................... 28
9. References ........................................................................................................................... 29
About the Author..................................................................................................................... 30
1. Executive Summary
When a hard disk drive containing valuable data no longer responds, the users last hope is to send the
drive to a data recovery company that specializes in drive hardware failures. There is a general perception
that data recovery companies have magic machines for retrieving data in almost any situation. The reality
is less glamorous. The most sophisticated, commercially successful recovery techniques involve careful
part-replacement, in a cleanroom environment, of the heads, the spindle motor and base casting, the
electronics board, and/or the drives firmware and parameter tables. Part-replacement has historically been
successful for data recovery about 40 to 60% of the time. Claimed data recovery success rates are much
higher. While they may, in fact, approach 100% for some drive models, for other models and failure modes
the success rate is near zero. Drive-independent data recovery methods are needed now to read these drives.
Furthermore, as the data density of hard disk drives continues to increase the number of unrecoverable
drives is expected to grow.
The reason for this lack of successful recovery can be traced to the methods drive manufacturers must
employ to achieve both high data density and high production yields. Specifically, current drives are
hyper-tuned in the factory to optimize the performance of each section of each hard disk drive. The data
format, head, disk, electronics, and firmware parameters are all optimized together. This means that it is
less likely that a head stack or electronics board or parameter tables from one drive even of the same
model will work well when used as a replacement in a failed drive.
ActionFront Data Recovery Labs SignalTraceTM technology is the only solution known to-date that
demonstrates the capabilities needed for commercially viable recovery of user data that is otherwise
unrecoverable using traditional part-replacement. SignalTraceTM technology replaces, instead, the exacting,
optimized signal processing and positioning functions of the disk drive with custom hardware, software,
and algorithms to precisely locate particular sectors of data and recover each bit individually independent
of the drives specific hardware. Furthermore, its underlying design has the flexibility to provide this data
recovery capability into the future as increasing data densities continue to require more hyper-tuning of
disk drives in the factory.
1000
(about 30 nanometers)1. The track-to-track spacing
(track pitch) is 8 microinches (about 200 nanometers).
100 The 5 10% guardbands between tracks are a fraction
of a microinch (less than 20 nanometers). It is
10
astonishing that drives routinely achieve this level of
1
mechanical precision at a price per megabyte that has
$/MB Paper or Film Storage been falling at the rates shown in the graph to the left.
0.1 For the past few years, it has been cheaper to store
2.5
data on HDDs than on paper or film. Currently the
0.01
price of HDD storage is about $1/gigabyte.
3.5
0.001
When drives cost thousands of dollars, drive repair
0.0001 was a lower priced alternative to purchasing a new
1980 1985 1990 1995 2000 2005
HDD. Today, the most economical option for dealing
Production Year with a malfunctioning drive is to replace it with a new
one. The new drive will likely be larger, cheaper, and
faster. In fact, it is typically the data itself even for the home user that is
much more valuable than the drive.
Increasingly, the home users drive is filled with often-priceless photos and
movies. The time it takes to recover a failed drive can also be more costly than
the drive itself even when backups are available. (You do have backups, dont
you?) However, backups typically represent a snapshot of the data some time
ago (last night, last week, last month). Therefore all recent work and
transactions are still lost. Unfortunately, many companies that run backups
diligently do not practice restoring data from backups. Sometimes the backups
themselves are corrupted. Even in redundant systems, such as drive arrays, data
loss due to multiple-drive failures is not uncommon.
For these reasons, no matter what precautions have been taken, a drive may need
the services of a data recovery company. For criminal investigations requiring
data forensic analysis, there is no substitute for the drive in question. It must
yield its information even if it has been intentionally destroyed.
1
A microinch is one millionth of an inch; a nanometer is one billionth of a meter. There are about 25.4
nanometers in 1microinch.
How does the drive know where your file is? It doesnt. That is the job of the
operating system. The operating system keeps track of which logical blocks on
which drive contain your file. For convenience, we will consider a logical block
The HDD does
to be a data sector, although each block could also point to several consecutive
not know where sectors. The drive will request a logical block from the drive, for example block
files are located. # 1,635,324. The HDD must map this logical block location into a physical
It knows where block (CHS) location, for example cylinder 5,000 on head 1 at sector 452. There
sectors are are fast algorithms for computing this, however the interesting complication is
when the usual physical location for a logical block has a defect that precludes it
located.
from reliably storing data.
Such locations are found and mapped out during the manufacturing process.
There are also provisions for doing this check and re-mapping when the drive is
in use in the field. The drive has many spare sectors and even spare tracks to be
used as replacements for defective sectors. This is transparent to the operating
system under normal operation. The drive accepts the logical block address and
performs the logical-to-physical translation itself. This varies from drive-to-
drive, reflecting the mapping-out of defects found during the drives surface
scan self-test.
In the field, the drive may acquire additional defects due to corrosion, handling,
or other causes. These are typically identified in a table of exceptions
(sometimes called the P-list and the G-list, for primary defects and grown
defects, respectively). This table, the table of parameters, and the
firmware are typically stored on the disk itself in the outermost
tracks. These tracks are referred to as the system area, maintenance
tracks, diskware, negative cylinders, etc. However, some drive
models store the table in non-volatile memory on the printed
circuit board. Clearly this table of exceptions is uniquely linked
to the media in a particular drive. The table for one drive will
not, in general, be the same for the media from another drive.
ID Radius OD To maximize the amount of data that can be stored, each disk
surface is divided into groups of adjacent tracks called zones. There are 8 to 32
(or more) zones per surface. From the ID to the OD, each zone is written with a
The servo pattern is typically written at a much lower bpi than the data and its
Track Motion,
frequency is constant across the disk. It is not zoned. This means that the bpi is
Relative to Head lower at OD. In other words, the servo pattern is shorter near the ID and longer
near the OD a wedge shape. There are typically 50 to 200 evenly spaced servo
wedges per revolution. This embedded servo
information is on each disk surface.
The first two bursts, typically called the A burst and the B burst, are shown
written off center. When the head is exactly on track center, it will get a certain
The servo system also identifies each sector. It does this by maintaining
synchronization with the first servo wedge in a revolution and timing from
there to indicate the beginning and end of each data sector on the track. This
timing relationship changes from zone-to-zone, but the wedge-to-wedge servo
timing remains constant.
For good error rate performance, it is necessary to establish the proper gain for
each sector and lock the detection process to the precise frequency and phase of
the readback waveform. This places three specific requirements on the stored
data.
1) Every data sector must start with a single-frequency sequence of
transitions. This is usually called the preamble and is about 10 to 15
bytes long. The preamble makes it much easier to establish the proper
gain and timing synchronization for the sector. Every servo field also
starts with a single frequency preamble for the same reason.
The PRML detection techniques require a target for the expected pulse shape
and for how pulses interfere with each other. To ensure that the waveform is
close to this target, a combination of fixed and adaptive filtering is applied to the
readback signal. For
best performance, all
of these channel
parameters must be
optimized (tuned)
for each zone of each
head in each drive.
ChannelSciences read
channel simulation
software package,
PRMLproTM (shown in
the figure to the left),
models most of the
signal processing used
for detecting the
sequences of 1s and 0s
from captured
readback waveforms
from magnetic disk,
tape, and optical drives
[4].
Even with all of these steps, the post-detection raw error rate is only about 10-5
to 10-8. In order to achieve the specified unrecoverable error rates of 10-13 to
10-15, error correction coding must also be used.
Because the bits are flipped pseudo-randomly, the flipping sequence can be
regenerated during readback so that the data is exactly unscrambled. Precise
location of the sync mark is necessary for this to succeed. Notice that the
scrambler does not prohibit any pattern. For example, it is possible for the user
to store a bit sequence that is scrambled into an all-zeros pattern. For this
reason, it is still necessary to apply an RLL code to the scrambled user data.
A common RLL code for PRML channels maps 16 scrambled data bits into 17
code bits. This is a coding overhead of about 6% (17/16). This type of code
ensures that there are no more than a certain number of zeros (maybe 10 to 15)
in between ones. This causes pulses to be present in the readback waveform
often enough for gain and timing to be tracked. There are other RLL codes that
have much higher rates than 16/17. There are also RLL codes that are designed
to eliminate certain patterns that are more error-prone. It is possible that
different RLL codes are used in different zones of a single disk surface.
Currently, most drives combine RLL codes with a parity check code. This
typically adds one or two bits to the RLL code overhead. For example, a 64/65-
rate code (64 user bits are encoded into 65 RLL code bits) would become a
64/66-rate code when a single parity-check bit is added. The benefit of adding
this small amount of parity is that the dominant errors made by the detector can
be identified and corrected with a small increase in circuitry and code overhead.
However, all of these encoding methods combined still do not achieve the
There is a very, unrecoverable read error rate goal of better than 10-13. This is possible only with
very small error correction coding (ECC). ECC calculates parity bytes for the users data,
chance that the which provide structured redundancy that can be used during decoding to detect
and correct errors. The ECC encoded user data is what is scrambled and RLL
users data will encoded. Typically, Reed-Solomon encoding is used because of its good burst
be returned error correction capability and the economy of its implementation. Bursts of
incorrectly. errors occur because a scratch or other small mark corrupts a group of
consecutive bits. It is not uncommon to have the ECC capability to correct over
200 bit errors in a sector.
The ECC can fail in two ways. One way is that there are too many errors in a
sector to correct. This is an unrecoverable read error. However, the drive will
If there are a few more errors in a sector than the ECC can correct, and they
occur in a certain way, it is possible that the ECC decoding miscorrects the data.
This is disastrous in financial transactions, for example. The probability of
miscorrection, also called the probability of data corruption, is not commonly
specified on drive data sheets. Ideally the probability is much less than 10-20. To
ensure that it is very unlikely that data will be miscorrected, the ECC encoded
data is often wrapped with a CRC (cyclic redundancy check) code. This has a
very strong capability to detect errors, but is not used for correction. This
provides the final check that the data is correct as delivered back to the computer
over the interface.
The figure below shows the encoding sequence and the organization of sectors
on a track. Notice that to get the most benefit from zoning, sometimes data
sectors are split across servo wedges. The second part of a split sector must
also start with a preamble and a sync mark. The detected data sequences from
both portions are concatenated and the decoding and descrambling proceed as
usual.
Servo Servo
Wedge Wedge
Data Sector Data Sector Data Sector Split begin Split end Data Sector
Inter-sector Gaps
After the bpi/tpi, zoning, writing parameters, and reading parameters are
determined, the detection parameters are optimized. These must be determined
for every zone of every surface of every drive. A 6-surface drive with 16 zones
requires 96 groups of channel optimization settings to be stored in the
parameters table. These channel settings include equalization and noise-
whitening filter coefficients; gain, timing, and adaptation parameters; detection
target; RLL code selection; etc. Similar settings must also be stored for detecting
the servo wedge information.
With almost every new generation, a drive parameter that was fixed becomes
variable. This new variable must then be optimized, which leads to the hyper-
tuning that occurs routinely in modern disk drives.
These include:
Failure of solder traces, electronic components, or connectors on the printed
circuit board (PCB)
Exceeding a S.M.A.R.T. (Self-Monitoring, Analysis, and Reporting
Technology) threshold
Damaged or corrupted firmware
Uncorrected bug in factory firmware
Damage to system areas of the disk that are used for calibration, testing,
storing firmware and parameters tables
Spindle or voice-coil motor failure (e.g., short circuit, open circuit)
Seized bearings
Breakdown of bearing grease
Disk shift or mis-alignment
Head damage
Overheating
And many others
A reputable data recovery company will tell you whether or not they have a
good success rate with your drive. However, they might not have received a
particular model yet for recovery and will not know for certain how likely a
recovery is. Furthermore, even if the success rate is good for a drive, your drive
may be damaged in such a way that recovery is not possible. There is always a
chance that the data cannot be recovered. When your critical data is on the line,
An independent you want to be sure that the data is unrecoverable because of the drive and not
data recovery because of the lack of skill at the data recovery company that you chose.
trade organization
would benefit the In the past, some drive manufacturers have had qualification programs in which
end-user and the they identified approved data recovery companies. Such programs appear to
have been dropped, perhaps because of the dangers of being implicated in
industry. lawsuits if the recovery fails or even makes things worse. On the websites of
several major drive manufacturers, their help for data recovery now is to
It would be helpful to the end-user, and to the most reputable and capable
companies in the industry, for an independent data recovery trade association to
be formed. This association could provide a certification program for data
recovery specialists and it might also gather and publish statistics on success
rates for recoveries on different models of drives. It could also certify
individuals in chain-of-evidence procedures for data forensic investigations. In
addition, it could direct military, law enforcement, and intelligence agencies to
the companies that have the highest success rates for particular types of
intentional drive damage.
An easier step toward transparency is for the best data recovery companies to set
the standard for the industry by example. This might include providing more
information about actual recovery techniques and the true nature of most
recoveries. It might also include listing success rates by model or by failure type
on their websites. However, unless this is adopted by the top companies at the
same time, such disclosures can give a negative impression to the end-user who
Submit claims of is searching for a data recovery company for the first (and only) time. They are
extraordinary likely to feel more comfortable with the company claiming to have magic
data recovery machines and proprietary processes that yield a success rate of over 90%.
capabilities to As a first step for the industry, all data recovery companies that make a claim of
the peer- special machines or proprietary processes that go beyond the standard
reviewed replacement of failed parts should present their unique capabilities for
scrutiny of the independent review. A straightforward way to accomplish this is to submit a
IEEE Transactions paper to the most important journal of peer-reviewed technical papers for the
disk drive industry the IEEE Transactions on Magnetics [7]. I call on all data
on Magnetics. recovery companies to submit their claims of extraordinary capabilities to the
scrutiny of this refereed journal.
This method can fail if the drive stores the parameters tables on non-volative
memory on the PCB. Howver, it is possible in some cases to transplant this
memeory chip
from the failed
drives PCB to the
donor PCB. This
method can fail if
the donor drives
PCB does not
Damaged PCB contain a very
(Printed Circuit Board) similar version of
the PRML read
channel to the one
on the failed drive.
This is because the
HDA channel settings in
(Head/Disk Assembly) the parameters
table might not
work for the new
chip. This method
can also fail if
there is additional damage to the drive that prevents it from seeking to the
system area and reading the drives firmware and parameter tables.
Clearly, anything that stops the drive from reading the firmware complicates
part-replacement considerably. If the media near the landing zone is damaged,
the servo might not be acquired. If
Landing zone there is a scratch at the OD that
damage destroys the system area, the
parameters for the hyper-tuned
drive are lost, along with the
firmware and defect management
information.
For either method of getting the firmware loaded into the failed drive, the best
outcome is that the drive will spin up, its own servo will be re-synchronized, and
seeking will be possible. Of course, the wrong defects will probably be
mapped out because the defect table from the donor drive will be in RAM.
System area
damage
Equal or greater care is necessary when loading the good heads back onto the
damaged drives media.
Once the good heads are loaded onto the remounted disks, the power-on
procedure can begin.
However, there are very special machines used by drive manufacturers for the
design and analysis of drive components. It is often suggested that these
precision instruments, spin-stand testers and magnetic force microscopes
(MFMs), can be used for data recovery.
Spin-stands are very accurate and flexible for analyzing raw disks. Virtually
any data pattern can be written and the positioning accuracy and repeatability
Spin-stand are in the nanometer range. However, this typically requires that the tester write
testers are its own servo pattern. Reading a disk that has been written by a drive is more
accurate, flexible problematic.
instruments that
illustrate the First the disk and head must be aligned as close to their relationship in the disk
drive as possible. Then the electronics and software must be programmed to
benefits of drive- utilize the servo pattern written on the disk. If the servo can be followed, the
independent test parameters for the head and channel still need to be optimized. Assuming that is
equipment. possible, the data written to the disk should be readable.
However, unless the exact read channel and its coding options are available for
the tester, all that will be delivered is scrambled, RLL encoded, ECC code words
at best. These must still be decoded and then assembled into useful files. Note
also that the head will be flying over the disk surface, so the disk must not be
significantly damaged.
In reality, the scenario above is very difficult to successfully implement even for
a drive manufacturer. It takes a great deal of trial-and-error investigation by a
very knowledgeable operator. It would be much more difficult for a data
recovery company to implement this technique successfully across virtually all
manufacturers drives cost-effectively.
The figure to the left is an MFM image of a portion of a track of data. The dark
and light horizontal lines are the individual transitions. Assuming the transitions
are 1s, the spaces in between the transitions are the 0s. The detail clearly reveals
1/bpi the guardbands between tracks and even the curl at the edges of the written track
due to the shape of the write field.
The MFM probe must be very close to the disk surface in order to get these
1/tpi images. Therefore it cannot easily follow a badly damaged (e.g., bent) disk. The
biggest drawback, however, is its speed. The MFM scans about a 100 micron by
100 micron area at a time, then the sample must be moved and the next area
scanned.
The most intriguing possibility for magnetic force microscopy as a data recovery
tool is reading overwritten data [11]. As shown in the
Track N Track N+1 image to the left, when a track is overwritten there is often
a portion of the previously written data remaining. This is
due to small variations in the servos placement of the
write element as well as the effects of spindle runout. It is
theoretically possible to take all the steps listed above but
generate the readback signal from in between tracks rather
than from track center. This procedure will have about the
same level of difficulty, but the error rate of the readback
signal will be much worse. Also the overwritten signal
Previously written will be slowly fading in and out due to non-repeatable
(partially overwritten) spindle runout that occurs during writing. Such an effort
data in guardband could only be afforded for a small amount of the most
important data for national security.
However, it still leaves all the problems of analyzing (quickly) the many
terabytes of image data generated. The images must be arranged in the correct
spatial pattern and the tracks followed by some image processing servo routine.
The readback signal from the track center (or guardband) must be generated.
And finally the data must be detected, decoded, and assembled into useful files.
An improvement on this system would be to servo the imaging head during the
scan by using the magnetic patterns written on the disk.
Because of this, firmware replacement is likely to fail for one or more of the
following reasons.
Channel settings for servo signal detection are too far off for good servo
reading
Servo synchronization is achieved but the head offset measurements are too
far off to yield proper seeking
The zone table information that identifies the layout of the drive and the
adaptive format information, such as bpi and tpi, is completely different
from the drive and no data can be read
Sectors needed for crucial files are listed as defective (from the donor
drives G-list and P-list) and defective sectors are listed as good
The channel settings for data are too far off to get good read error rate
performance
To make this possible, certain key physical parameters of the drive are
measured, or calibrated, in the factory. For example, one such parameter is the
offset between the read and write elements on each head, and how the
relationship changes from track-to-track due to the effects of skew angle.
Another parameter is related to the fact that the tracks are not perfect circles.
This is called eccentricity and its effect is referred to as repeatable runout
(RRO). RRO can be measured on each surface and a periodic term added to the
servo algorithm to compensate for this predicable movement of the track relative
to the head during each revolution. For clarity, the figure to the left illustrates
RRO caused by a shift in the disk center relative to the location of the pivot
point for the new replacement head stack.
If this happens in the field (perhaps due to a rotational shock), the worst
consequence is that a write will be executed, based on the wrong servo timing
when a servo wedge is under the write element. This write operation will destroy
the servo information. If this continues for even a few wedges, the surface (and
hence the drive) is likely to be unreadable by normal means.
HDD manufacturers typically use carefully controlled robots to place the disks
and spacers in the pack, balance the disks (if needed), and torque the retaining
screws precisely. The disk can warp (or potato chip) if retaining screws do not
provide even pressure or if a spacer is not very flat. The head can follow a
certain amount of warpage as it flies, but it is possible to have excessive potato
chipping that results in erratic flying. It is also possible that the disks motion
and the windage that it generates can excite certain mechanical resonances in the
suspension and arm, which can make precise servoing very difficult.
Disk remounting is likely to fail for one or more of the following reasons.
Disk slip
The centers of the disks line up differently, resulting in disk-to-disk
eccentricities that are different than the servo is programmed to correct
Disk warp
The spacing of the disks might be different relative to the spacing of the
heads in the head stack; this can result in load force differences that cause
excessive flying height differences
The spacing of the disks might be different relative to the spacing of the
heads in the head stack, which can make head reloading difficult, possibly
resulting in disk damage
The thermal stability of bits drops rapidly as the areal density increases there
are fewer atoms in each bit to retain the magnetic orientation. High temperature
environments can make thermal decay worse. It has also been demonstrated that
writing to a particular track can cause degradation in the bits in the adjacent
tracks. It is not widely known, but many modern drives routinely check for
thermal decay of bits in the field and rewrite the sectors in which degradation is
identified.
Unfortunately, there are likely to be many memories lost in the future as home
videos, long forgotten in a hot attic, are replayed only to find the image
degraded or lost due to thermal decay.
At this point, the prototype SignalTraceTM system is acquiring the servo wedges
and synchronizing to them. It commands the motor controller to make fine
SignalTraceTM adjustments to the RPM as needed, based on the servo wedge timing. It also
controls the drive finds the once-around spindle index provided by the servo. It does this with
whichever head is selected by the SignalTraceTM software in the control PC.
to seek to any
track on any From the PC, the drive can be commanded to seek to any track (with any head).
surface. The seek-and-settle time is 5 to 10 seconds. A single-track seek takes less than 1
second. The servo control algorithm is implemented in a Motorola MCS5407
ColdFire microprocessor. Note that the system must be pre-programmed with
the servo layout, the zone frequencies, the channel parameters, and the codes
used. These are determined off-line.
[B] Pre-determined
channel parameters
[G] Lastly,
errors can
be corrected
by the ECC
algorithm
The results for simulated continuous-time filtering are shown in [C]. The rest of
TM
SignalTrace has the channel signal processing is not shown explicitly. However, automatic gain
demonstrated the control (AGC), phase-locked loop (PLL) controlled sampling, and adaptive
finite impulse response (FIR) filtering are performed in the PC. The detection of
complete control 1s and 0s is performed and the results are displayed [D].
of a disk drive
and the The sync mark must be found in this sequence before RLL decoding can
successful proceed [E]. After the RLL decoding, the data must still be descrambled before
retrieval of user the ASCII text and unique LBA written by the utility program can be seen [F].
Note, the requested LBA was found, but that portion of the decoded sector is not
data. shown in the figure above. A few errors were made in the detection process (not
on purpose). These were corrected by the ECC [G].
This is a This demonstrates the complete control of a disk drive and the returning of
milestone for the corrected user data without relying on any electronics (except the preamp inside
the HDA) or signal processing from the drive itself.
data recovery
industry. ActionFront and ChannelScience worked together to overcome many long-
standing challenges in order to achieve this milestone in data recovery history.
An especially important advancement is the cryptographic procedures employed
by the research staff at ActionFront to descramble, RLL decode, and ECC
correct the raw detected data. This was reverse engineered, based on first-
principles analysis of a good drive of the same model. These highly specialized
techniques as well as the determination of many channel parameters, servo
layout and data layout must be applied to each new drive model before
recovery can be attempted. This is because the needed information for drive-
independent data recovery is not readily available from the drive and channel
companies.
To further document and verify this milestone for the data recovery industry,
ActionFront has agreed to submit a paper on SignalTraceTM technology to a
refereed technical journal.
SignalTraceTM Drives continue to evolve, getting more sophisticated, adaptive, and hyper-
Technology will tuned. For data recovery of hardware-failed drives to continue to be successful,
provide an drive-independent data recovery techniques, such as SignalTraceTM technology,
must be made commercially viable. Furthermore, they must work for most
important tool for popular drive models and they must continue to accommodate the relevant new
law-enforcement innovations in HDDs. An important additional benefit of drive-independent data
and counter- recovery is that it can be compatible with exotic data acquisition techniques for
terrorism retrieving readback signals from intentionally damaged disks. This can be a
professionals. significant tool for law enforcement and counter-terrorism professionals.
8. Conclusions
The majority of drives that are sent to data recovery companies for hardware
failure are a few years old. While some of them still respond well to traditional
Backup often part-replacement, there are some that are almost never recoverable. These may
and test the have been hyper-tuned in the factory so that high data density can be achieved
together with high manufacturing yields, and/or they may have corrupted system-
backups, or areas on the disk where drive parameters tables are stored. Such drives require a
very precise matching of the characteristics of the head, disk surface, and the
system parameters that is not possible with traditional part-replacement. As data
density continues its rapid increase, it is expected that fewer hardware-failed
drives will be recoverable with traditional part-replacement.
Therefore, it is likely that the capability to recover data from almost all of the
latest drives will only be available from the best of the best -- the data recovery
companies that other data recovery companies turn to for their most challenging
tasks. Drive manufacturers could help data recovery efforts by providing features
such as special commands to load and run optimization routines that allow part
replacement to work better. However this is unlikely given the effort that drive
companies must devote to increasing areal density, manufacturing yields, and
reliability.
ActionFront Data Recovery Labs Inc. is the first and only company to publicly
demonstrate the capability of drive-independent data recovery, with its
SignalTrace Technology. ChannelScience assisted with portions of the
development of this capability. SignalTrace Technology has demonstrated the
capabilities needed to recover data that is currently unrecoverable by traditional
part-replacement. Furthermore, the business goal is to make this technology-
intense method commercially viable, so that it is within reach of the individual
end-user of hard disk drives not exclusively large corporations and government
agencies.
9. References
[1] Hitachi Global Storage Technologies website, www.hgst.com, follow links to
About Us, Company Background, Technology Timeline.
[2] R.A. Cideciyan, et al., A PRML System for Digital Magnetic Recording,
IEEE Journal on Selected Areas in Communications, January 1992, pp. 38-56.
[3] R.A. Cideciyan, et al., Noise Predictive Maximum Likelihood Detection
Combined with Parity-Based Post-Processing, IEEE Transactions on
Magnetics, March 2001, pp. 714-720.
[4] A 30-day Free Trial version of PRMLproTM can be downloaded from
www.ChannelScience.com.
[5] Professional data recovery definition is from ActionFront Data Recovery Labs
Data Emergency Guide.
[6] J. Donovan and J. Kim of TrendFocus, Growth Returns to a Leaner, Healthier
HDD Industry, Insight, January-February, 2004. Available at www.idema.org.
[7] More information on the IEEE Transactions on Magnetics is available at
www.ieee.org.
[8] US Department of Defense Broad Agency Announcement (BAA), May 14,
2003:
https://www.bids.tswg.gov/tswg/bids.nsf/(BAAView)/24C936C9BBF67CCD85
256CDF007C8FC3/$FILE/DAAD_03_T_0024.pdf
[9] Guzik Technical Enterprises Inc. makes spin-stands that are very widely used
throughout the HDD industry, www.guzik.com
[10] Veeco Inc. makes popular MFMs and other tools for nanoscale metrology,
www.veeco.com
[11] R. Gomez, A. Adly, I. Mayergoyz, and E. Burke, "Magnetic Force Scanning
Tunnelling Microscope Imaging of Overwritten Data", IEEE Transactions on
Magnetics, September 1992, pp. 3141-3143.
[12] I.D.Mayergoyz, C. Serpico , C. Krafft, and C. Tse, Magnetic Imaging on a
Spin-Stand, Journal of Applied Physics, Vol. 87, No.9, May 1, 2000, pp.6824-
6826.
Chuck is Chief Scientist of ChannelScience and the creator of PRMLproTM, a commercial software tool
that replicates most of the detection-related signal processing that occurs in magnetic disk, tape, and optical
data storage devices. His technical interests include nanotechnology-based data storage methods,
integrating hard disk drives into non-traditional applications, physiological monitoring, and applying
advanced adaptive data detection techniques to chemical and biological warfare sensors for homeland
security.
A favorite challenge is helping companies prepare their new technologies for acceptance in the hard disk
drive industry. Chuck can also be found on the other side of the table, evaluating the technologies of
companies that are investment candidates or merger and acquisition targets. He also assists large, small,
and emerging companies with technology roadmap assessment and technical marketing.
Chuck has five issued US patents and several published papers and articles. He has authored and taught
very popular seminars around the world for data storage professionals through KnowledgeTek Inc. These
include such topics as hard disk drive technology, servo positioning, PRML detection, error correction
coding (ECC) and iterative detection. A free download of the 30-day trial version of PRMLproTM is
available at www.ChannelScience.com.
Please Note
This white paper is for information only and is provided as is. The reader is encouraged to refer to original
sources and to check the patent ownership of ideas or technology presented herein before using any of the
information provided. Not all data recoveries are successful, regardless of the company or method used.
The author is not responsible for any direct or consequential losses due to the use of the information
contained herein.